JP4211023B2 - Moving image processing method and moving image processing apparatus - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention sends a bit stream obtained by compressing and encoding an input moving image to the receiving side by wire or wirelessly, or records it on a recording medium such as a magnetic disk or an optical disk, receives the bit stream, decodes it, or The present invention relates to a moving image processing method and a moving image processing apparatus for displaying a moving image by reproducing a bit stream from a recording medium.
[0002]
[Prior art]
Various methods are already known for compressing and encoding moving images. For example, MPEG-1, 2, 4 and the like as MPEG (Moving Picture Experts Group) systems have been internationally standardized. This MPEG system divides a screen into blocks composed of a plurality of pixels, and transmits or stores a bit stream encoded by performing compression processing in a spatial direction and a time direction in units of blocks. Is restored to the original moving image and displayed on the display.
[0003]
The above-described spatial compression in the compression encoding of a moving image is a process using correlation within a frame. In the case of the above-described MPEG system, orthogonal transformation by DCT (Discrete Cosine Transform), quantization, This includes processing with variable length coding. For example, the screen is encoded in units of macroblocks (MB) made up of 16 × 16 pixels. In the DCT process, two-dimensional DCT is performed on an 8 × 8 pixel block obtained by dividing MB into four.
[0004]
By this DCT processing, a DCT coefficient biased toward low frequency components is obtained, and 0 is increased for high frequency components. The DCT coefficients are quantized with a predetermined quantization scale, and the quantized coefficients are encoded using a run-length two-dimensional variable length code. In this case, since the quantization coefficient is also biased toward the low frequency component, variable length coding is performed by sequentially scanning from the low frequency component to the high frequency component side. The above-described compression processing in the spatial direction is called intra-frame coding.
[0005]
The compression in the time direction compresses generated information by applying inter-frame motion prediction. For example, the quantization coefficient obtained by the quantization process before the variable length coding process is inversely quantized, and the DCT process is performed. The difference between the locally decoded image decoded by performing the reverse DCT process and the input moving image is obtained. That is, an inter-frame difference is obtained, and the difference information is encoded by the above-described DCT processing, quantization processing, and variable length encoding processing to form a bit stream. Furthermore, a motion search process is performed to search from which position of the already-encoded image (local decoded image) each block has moved between frames, a position with a small inter-frame difference is obtained, a motion vector by the motion search process, The difference block from that position is encoded.
[0006]
In the decoding process, for intra-frame coding, the variable length decoding unit decodes the variable length code to restore the quantization coefficient, the inverse quantization unit restores the DCT coefficient, and the inverse DCT unit performs the inverse DCT. Processing is performed to restore the frame data. For interframe coding, variable length decoding, inverse quantization, and inverse DCT are used to restore the difference information of each block, and the frame data blocks that have already been decoded and stored in the frame memory. The moving image can be restored by adding the block at the position indicated by the motion vector.
[0007]
[Problems to be solved by the invention]
Since the conventional compression encoding process of moving images uses the fact that the correlation between the spatial direction and the temporal direction of the moving images is large, compression is not applied to moving images that do not have such properties. Efficiency decreases and the amount of generated information increases. In order not to decrease the compression efficiency, control such as increasing the quantization step is required, but this causes a problem that the reproduction image quality deteriorates and an undesirable display image quality is obtained. In addition, the amount of generated information increases due to a decrease in compression efficiency, and even if control of the quantization step to avoid it is performed, there may be a situation where it is not in time, in which case compression encoding processing is not performed. There is a problem that hangs up.
[0008]
For example, in the case of a moving image that picks up the rippled water surface irradiated with intense light, or in the case of random noise such as snow noise that results in a playback screen that snows, correlation within the frame and correlation between frames In other words, the correlation between the spatial direction and the time direction becomes small. Therefore, the compression efficiency is reduced, the amount of generated information is increased, and the overflow of the transmission buffer is generated. Therefore, it is necessary to control the amount of generated information to be reduced by increasing the quantization step. Also, a noise generation screen such as snow noise is an undesirable screen.
It is an object of the present invention to control so as to avoid a reduction in compression efficiency and to prevent an undesirable reproduction screen.
[0009]
[Means for Solving the Problems]
  Referring to FIG. 1, the moving image processing method of the present invention is a moving image processing method for transmitting or accumulating a bit stream obtained by compressing and encoding an input moving image.Obtained from at least one of statistical information on frame luminance in the compression encoding process, motion vector statistical information on motion vectors, and statistical information on the amount of generated coding information or quantized values.StatisticsAnd the judgment valueInput videoButSpecific sceneor notIs determined, and when it is determined as a specific scene,A unit in which the bit stream of this specific scene is formed as one stream, for example, a GOP (Group of Picture) unit,Pre-stored bit streamSwitch to the selector section 5 etc.It includes a process of outputting.
[0010]
In addition, at least one of the horizontal component average, horizontal component variance, vertical component average, and vertical component variance of the motion vector of the statistical information in the compression coding process of the input moving image is obtained and compared with a predetermined judgment value. In addition, a process of determining the specific scene when the determination value is exceeded may be included.
[0011]
The moving image processing apparatus according to the present invention is a moving image processing apparatus that transmits or stores a bit stream obtained by compressing and encoding an input moving image, the encoding unit 1 that compresses and encodes the input image, and the input moving image. The statistical information acquisition unit 2 that acquires the statistical information or the statistical information in the compression encoding process and the statistical information acquired by the statistical information acquisition unit 2 are compared with the determination value to determine whether the scene is a specific scene. A bit stream memory 4 for a bit stream memory 4 that stores a predetermined bit stream, a bit stream memory 4 that stores a predetermined bit stream, and a bit stream that is compression-encoded by the encoding unit 1 when the scene determination unit 3 determines that the scene is a specific scene. And a selector unit 5 for switching and sending a predetermined bit stream stored in the.
[0012]
The statistical information acquisition unit 2 is a statistical information acquisition unit that calculates an average frame luminance of an input moving image, or a statistical information acquisition unit that calculates horizontal and vertical component averages and variances of motion vectors in the encoding unit 1 or an average quantization scale. It can be configured to include at least one of statistical information acquisition units for obtaining values and generated information amounts. The bit stream memory 4 may be configured to store a predetermined bit stream corresponding to a bit stream or compression encoding a still image or a moving image with slow motion.
[0013]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 is a diagram for explaining the principle of the present invention. An encoding unit 1 performs compression encoding processing using correlation between a spatial direction and a temporal direction for an input moving image, and applies the above-described MPEG method. Can do. The statistical information acquisition unit 2 is a statistical information acquisition unit that calculates the frame luminance average of the input moving image, the statistical information acquisition unit that calculates the horizontal and vertical component average and variance of the motion vector in the encoding unit 1, or the average quantum It includes at least one of a statistical information acquisition unit for obtaining a scale value and generation information.
[0014]
The scene determination unit 3 compares the statistical information acquired by the statistical information acquisition unit 2 with the determination value to determine whether the scene is a specific scene, and controls the selector unit 5. The bit stream memory 4 stores a predetermined bit stream corresponding to, for example, a bit stream obtained by compressing and encoding a still image or a moving image with a slow motion. Then, the selector unit 5 switches between the bit stream from the encoding unit 1 and the bit stream from the bit stream memory 4 according to the determination of the scene determination unit 3 and sends it out.
[0015]
For example, since the compression efficiency is reduced for the above-described picked-up water surface image irradiated with intense light, the scene determination unit 3 determines a specific scene based on statistical information. A bit stream obtained by compressing and encoding the still image is repeatedly read out from the bit stream memory 4 and transmitted. Thereby, a decrease in compression efficiency in the compression encoding process is not a problem, and the playback display screen has stable display contents. It is also possible to store a plurality of types of bitstreams in the bitstream memory 4 and control to send one bitstream selected from the plurality of types according to the determination result of the scene determination unit 3.
[0016]
FIG. 2 is an explanatory diagram of an embodiment of the present invention, in which 11 is a frame memory, 12 is an input moving image statistical information acquisition unit, 13 is a motion vector statistical information acquisition unit, 14 is an encoded statistical information acquisition unit, and 15 is 16 is a bit stream memory, 17 is a selector, 18 is an original MB unit, 19 is a reference block unit, 20 is a motion vector searcher, 21 is a prediction determiner, 22 is a subtractor, 23 and 25 are selectors, 24 is an adder, 26 is a discrete cosine transformer (DCT), 27 is a quantizer (Q), 28 is an inverse quantizer (IQ), 29 is an inverse discrete cosine transformer (IDCT), and 30 is a variable length code. Figure 2 shows a generator (VLC).
[0017]
The statistical information acquisition unit 2 in FIG. 1 includes at least one of the input moving image statistical information acquisition unit 12, the motion vector statistical information acquisition unit 13, and the encoded statistical information acquisition unit 14 in FIG. Therefore, it is possible to provide all of them. 2 corresponds to the scene determination unit 3 in FIG. 1, the bit stream memory 16 in FIG. 2 corresponds to the bit stream memory 4 in FIG. 1, and the selector 17 in FIG. This corresponds to the selector unit 5 in FIG.
[0018]
1 includes the frame memory 11, the discrete cosine transformer 26, the quantizer 27, the inverse quantizer 28, the inverse discrete cosine transformer 29, the variable length encoder 30 and the like shown in FIG. It has the structure containing. The frame memory 11 stores an input moving image and stores a locally decoded image. An original image MB unit 18 inputs a macro block (MB) of the input moving image to the motion vector searcher 20 to perform local decoding. Reference blocks are sequentially extracted from the image by the reference block unit 19 and input to the motion vector searcher 20 to obtain a motion vector.
[0019]
The state in which the selector 23 selects the subtractor 22 side as shown in the figure and the selector 25 selects the adder 24 side as shown in the figure is the case of inter-frame coding, and the selectors 23 and 25 do the opposite. The state in which the side is selected indicates the case of intra-frame coding. Then, it is orthogonally transformed by the discrete cosine transformer 26, quantized by the quantizer 27, converted to a variable length code by the variable length encoder 30, and transmitted or accumulated via the selector 17. In addition, the quantized coefficient by the quantizer 27 is converted back to the DCT coefficient by the inverse quantizer 28, decoded by the inverse discrete cosine transformer 29, and temporarily stored in the frame memory 11 as a locally decoded image in the case of intra-frame coding. In addition, in the case of inter-frame coding, the adder 24 adds the block via the prediction determiner 21 to obtain a locally decoded image.
[0020]
In the input moving image statistical information acquisition unit 12, the frame luminance average is obtained. If the set of pixels in the frame is U, the average of frame luminance is AveY, the number of pixels is Num, and the luminance of the pixels is Pixel,
AveY = (Σ_UPixel i) / Num i (1)
It can be expressed as.
Also, the frame luminance variance VarY is
VarY = (Σ_U(| Pixel i-A |)) / Num i (2)
It can be expressed as. Note that the cumulative addition Σ in equations (1) and (2)_UThe symbol of indicates that i⊂U is added.
[0021]
The motion vector statistical information acquisition unit 13 accumulates the motion vectors obtained for each MB and divides the accumulated value by the number of MBs of the picture to obtain an average of the motion vectors. The motion vector variance can be obtained by calculating the sum of squares of the difference between and the average of the motion vectors and dividing by the number of MBs. In this case, the MB set in the frame is V, and the horizontal component of the motion vector is VecH. i, the vertical component of the motion vector is VecV As i, the horizontal component average AveHV, horizontal component variance VerHV, vertical component average AveVV, and vertical component variance VerVV of the motion vector are:
AveHV = (Σ_VVecH i) / Num i (3)
VerHV = (Σ_V｜ VecH i-AH |) / Num i (4)
AveVV = (Σ_VVecV i) / Num i (5)
VerVV = (Σ_V｜ VecV i-AV |) / Num i (6)
It can be expressed as. In addition, Σ of cumulative addition in equations (3) to (6)_VThe symbol of indicates that i⊂V is added.
[0022]
In the encoded statistical information acquisition unit 14, the amount of information generated in the encoding processing unit including the discrete cosine transformer 26, the quantizer 27, and the variable length encoder 30, the average of quantized values, etc. Can be obtained as encoded statistical information. For example, the MB set in the frame is V, and the amount of generated information of each MB is Bit. i, the amount of generated information of a picture is SumB, and the quantization scale value of each MB is Q i, If the average quantization value is AveQ,
SumB = Σ_V(Bit i) (7)
AveQ = (Σ_VQ i) / Num i (8)
It can be expressed as. Note that Σ of cumulative addition in equations (7) and (8)_VThe symbol of indicates that i⊂V is added.
[0023]
In the scene determination unit 15, the above-described statistical information and the determination values α, β, χ, obtained by the input moving image statistical information acquisition unit 12, the motion vector statistical information acquisition unit 13, and the encoded statistical information acquisition unit 14 are obtained. A specific scene is determined by comparing δ and ε. That is,
VarY> α (9)
AveHV> β (10)
VarHV> χ (11)
AveVV> δ (12)
VarHV> ε (13)
Can be determined. When any one of the conditions (9) to (13) is satisfied, or when a plurality of conditions are satisfied, it is determined that the scene is a specific scene when the statistical information satisfies a condition larger than the determination value. The scene determiner 15 sets such determination logic.
[0024]
The switching of the bit stream in the selector 17 by the control signal from the scene determiner 15 can be considered as a unit that is established as one stream in consideration of the inter-frame difference. For example, in the case of MPEG-2, a GOP (Group of Picture) unit can be used. Even when a predetermined bit stream read from the bit stream memory 16 is selected and transmitted, the compression encoding process is continued, and the input moving image is compressed according to the determination result of the scene determination unit 15. Control is performed so that a bit stream by encoding can be switched and transmitted.
[0025]
FIG. 3 is a flowchart according to the embodiment of the present invention. Input video is read (A1), header generation (A2), motion search (A3), MB coding (A4) is performed, and whether or not the picture is End. (A5), if compression encoding in units of pictures is continuing, the process proceeds to step (A3), and if completed, essential information is acquired (A6), and it is determined whether the determination condition is satisfied. (A7).
[0026]
Also, the input moving image is read, the input moving image statistical information acquisition unit 12 acquires the statistical information such as the frame luminance average (A11), and the motion vector search unit 20 performs the motion search to acquire the motion vector statistical information. The statistical information such as the horizontal and vertical components of the motion vector is obtained by the device 13 (A12), and the amount of generated information and the average quantization value of the picture are obtained by the coded statistical information obtaining device 14 in the process of MB coding. And so on (A13).
[0027]
Such statistical information is acquired and compared with the determination value in the scene determination unit 15. If the above-described statistical information does not satisfy the condition larger than the determination value, a bit stream of the encoding result is output (A8). ), It is determined whether or not the encoding is finished (A9). If not finished, the process proceeds to step (A1), and if finished, the process is finished. In step (A7), when the scene determination unit 15 determines that the determination condition is satisfied, the selector 17 is controlled so as to output a predetermined bit stream stored in the bit stream memory 16.
[0028]
Therefore, when a bit stream is received, decoded and reproduced and displayed, a reproduction screen such as snow noise can be avoided and a stable display screen can be obtained. In addition, since the compression encoding process does not need to be transmitted when the amount of generated information is remarkably increased, it is not necessary to consider the overflow of the transmission buffer, etc., and stable compression encoding can be continued. It is possible to avoid an undesirable reproduction screen.
[0029]
Also, on the bitstream decoding side, it is determined whether or not the scene is a specific scene based on statistical information in the decoding process, and in the case of a specific scene, a predetermined bit that can constitute a stable playback screen The stream can be read from the bitstream memory and decoded. Thereby, it is possible to avoid an undesirable reproduction screen such as snow noise.
[0030]
FIG. 4 is an explanatory diagram of the decoding unit according to the embodiment of the present invention, which can be used in combination with the encoding unit shown in FIG. 2 or in combination with the conventional encoding unit. is there. In the figure, 41 is a decoder buffer, 42 is a bit stream memory, 43 is a selector, 44 is a decoding processor, 45 is a variable length decoder (IVLC), 46 is an inverse quantizer (IQ), 47. Is an inverse discrete cosine transformer (IDCT), 48 is a header information extraction unit, 49 is a header information acquisition unit, 50 is a decoding statistical information acquisition unit, 51 is a scene determination unit, 52 is a selector, 53 is an adder, and 54 is Indicates a frame memory.
[0031]
The selector 43 is controlled by the scene determiner 51 and normally selects the decoder buffer 41 side as shown. The bit stream memory 42 stores a single type or a plurality of selectable predetermined bit streams excluding header information. The bit stream that passes through the selector 43 is restored to the fixed-length code quantization coefficient by the variable-length decoder 45 of the decoding processing unit 44, restored to the original DCT coefficient by the inverse quantizer 46, and the inverse discrete cosine. When the converter 47 performs inverse DCT processing and decodes and decodes the inter-frame coding bit stream, the selector 52 is in the state shown in the figure, and the adder 53 performs block and block from the frame memory 54. When it is added and stored in the frame memory 54 and the intra-coded bit stream is being decoded, it is stored in the frame memory 54 without going through the adder 53. The reproduced image is displayed on the omitted display unit.
[0032]
The header information extraction unit 48 extracts the header added to the bit stream, identifies whether it is interframe coding or intraframe coding, and controls the selector 52. The decoding statistical information acquisition unit 50 acquires statistical information such as the horizontal and vertical components of the motion vector and the quantization value in the decoding processing unit 44, and the scene determination unit 51 determines the same as described above. If it is determined that the scene is a specific scene as compared with the value, the selector 43 is controlled to repeatedly read out the bit stream from the bit stream memory 42 and input it to the decoding processing unit 44. Thereby, a stable screen can be reproduced and displayed instead of an undesired screen such as snow noise.
[0033]
For streams such as digital broadcasting, a system stream is distributed as a single stream of programs of a plurality of channels, and a table to which information on each program is added is embedded in the stream. By reading this table, for example, violent programs and adult programs can be identified. The header information acquisition unit 49 acquires such header information, and for programs that the child does not want to show to the child, by setting the header information as a specific scene in the scene determination unit 51, an undesired screen is obtained. By controlling the selector 43 and selecting the bit stream memory 42 side, a predetermined bit stream can be decoded and reproduced and displayed.
[0034]
In this case, it can be a suitable image stream or a title image for prohibiting viewing of an adult program. That is, in the bit stream memory 42, an undesirable reproduction screen is displayed based on the statistical information acquired by the decoded statistical information acquisition unit 50, and an undesirable reproduction screen is generated by the header information acquired by the header information acquisition unit 49. It is possible to store so that different bitstreams can be selected in the case where it is determined that.
[0035]
FIG. 5 is a flowchart of the decoding process according to the embodiment of the present invention. The header of the bit stream is decoded (B1) and necessary information is acquired (B2). That is, information by statistical information acquisition by header decoding (B11) or information of the above-mentioned program type is acquired, and information by statistical information acquisition (B12) in the decoding process is acquired. Then, it is determined whether or not the determination condition is satisfied by the scene determination unit 51 (B3), and if it is determined as a specific scene such as an undesired reproduction screen such as snow noise or an undesired screen such as an adult program, a predetermined bit The stream is read from the bit stream memory 42 (B10), and when the determination condition is not satisfied, the received bit stream is read (B4), MB decoding (B5), and motion compensation (B6) are decoded. It is determined whether or not the picture is End (B7). If it is not finished, the process proceeds to Step (B5). If it is finished, the decoding result is output (B8), and it is determined whether or not the decoding process is finished ( B9) If not finished, the process proceeds to step (B1). If finished, the decoding process is finished.
[0036]
The present invention is not limited to the above-described embodiments, and various additions and modifications can be made. The encoding process can be an encoding process other than MPEG-2. .
[0037]
(Supplementary Note 1) In a moving image processing method for transmitting or storing a bit stream obtained by compressing and encoding an input moving image, the input is based on statistical information of the input moving image or statistical information in a compression encoding process. A moving image processing method comprising: determining a specific scene of a moving image, and selecting and outputting a predetermined bitstream stored in advance when the specific scene is determined.
(Supplementary Note 2) At least one of horizontal component average, horizontal component variance, vertical component average, and vertical component variance of the motion vector of the statistical information in the compression coding process of the input moving image is obtained and set in advance. 2. The moving image processing method according to claim 1, further comprising a step of comparing with a determination value and determining the specific scene when the determination value is exceeded.
[0038]
(Supplementary note 3) In a moving image processing apparatus for transmitting or storing a bit stream obtained by compressing and encoding an input moving image, an encoding unit for compressing and encoding the input image, and statistical information or compression of the input moving image A statistical information acquisition unit that acquires statistical information in the encoding process, a scene determination unit that compares the statistical information acquired by the statistical information acquisition unit with a determination value, and determines whether the scene is a specific scene; A bit stream memory for storing a bit stream and a predetermined bit stream stored in the bit stream memory for a bit stream compressed and encoded by the encoding unit when the scene determination unit determines that the scene is a specific scene. A moving image processing apparatus comprising a selector unit for sending out.
[0039]
(Supplementary Note 4) The statistical information acquisition unit is a statistical information acquisition unit that calculates a frame luminance average of the input moving image, or a statistical information acquisition unit that calculates horizontal and vertical component averages and variances of motion vectors in the encoding unit, or The moving image processing apparatus according to supplementary note 3, including at least one of statistical information acquisition units for obtaining an average quantization scale value and a generated information amount.
(Supplementary note 5) The moving picture according to supplementary note 3, wherein the bit stream memory has a configuration in which a predetermined bit stream corresponding to a bit stream obtained by compressing and encoding a still image or a moving image with moderate motion is stored. Image processing device.
[0040]
(Supplementary note 6) In a moving picture processing apparatus for decoding a compression-encoded bitstream, a bitstream memory storing a predetermined bitstream, and header information or decoding process of the compression-encoded bitstream A scene determination unit for acquiring statistical information in the process to determine whether or not the scene is a specific scene; and when the scene determination unit determines that the scene is a specific scene, a predetermined bit stream is read from the bit stream memory and the decoding process is performed A moving image processing apparatus, comprising: a selector that inputs to the unit.
(Supplementary note 7) The moving picture processing apparatus according to supplementary note 6, wherein the bit stream memory has a configuration in which a plurality of types of bit streams to be selected and read according to the determination content of the scene determination unit are stored.
[0041]
【The invention's effect】
As described above, according to the present invention, a predetermined bit stream is stored in the bit stream memories 16 and 42, and a scene determination device determines when an undesired reproduction screen is obtained based on statistical information and header information. The encoding side sends out a predetermined bit stream read from the bit stream memory, and the decoding side decodes the predetermined bit stream read out from the bit stream memory. Even if there is an increase in the amount of generated information in the case of a playback screen, since it is not transmitted, the compression encoding process can be continued stably, and there is an advantage that the playback screen also becomes a stable screen.
[Brief description of the drawings]
FIG. 1 is a diagram illustrating the principle of the present invention.
FIG. 2 is an explanatory diagram of an embodiment of the present invention.
FIG. 3 is a flowchart of an embodiment of the present invention.
FIG. 4 is an explanatory diagram of a decoding unit according to the embodiment of this invention.
FIG. 5 is a flowchart of decoding processing according to the embodiment of this invention.
[Explanation of symbols]
1 Encoding unit
2 statistical information acquisition part
3 Scene judgment part
4-bit stream memory
5 Selector section

Claims (4)

In a moving image processing method for generating and transmitting or storing a bit stream obtained by compressing and encoding an input moving image,
In the compression encoding process of the input video, statistics about the frame luminance, and statistical information obtained from at least one of the statistical information about statistics, coding amount of information generated or quantization values of the motion vectors comparing the judgment value, a process in which the input moving image to determine whether low made specific scene of compression efficiency,
Bits accumulated in advance in a unit that is formed as one bit stream including the specific scene with low compression efficiency among the bit streams generated in the compression encoding process determined as the specific scene with low compression efficiency A moving image processing method comprising a step of switching to a predetermined compression-encoded bit stream that does not reduce the compression efficiency of the stream and outputting the same. The statistical information about the motion vector in the compression coding process of the input moving image includes at least one of a horizontal component average, a horizontal component variance, a vertical component average, and a vertical component variance of the motion vector. hints, the comparison stats and preset determination value, the statistical information is the determination value moving picture processing according to claim 1, wherein said to contain a specific scene and determines the course when exceeded Method. In a moving image processing apparatus for transmitting or storing a bit stream obtained by compressing and encoding an input moving image,
An encoding unit for compressing and encoding the input moving image,
And statistical information about the in-frame luminance compression encoding process by the coding unit, wherein the statistical information in the motion vector to the compression encoding process, the compression encoding process in the encoding generated information amount or A statistical information acquisition unit for obtaining statistical information from at least one of the statistical information about the quantized value;
A scene determination unit that compares the statistical information acquired by the statistical information acquisition unit with a determination value, and determines whether or not the specific scene has a reduced compression rate of the input moving image ;
A bit stream memory for storing a predetermined compression-encoded bit stream that does not reduce the compression efficiency of the bit stream;
One stream including a specific scene having a low compression efficiency among the bitstreams compression-encoded by the encoding unit when the scene determination unit determines that the specific scene has a reduced compression rate of the input moving image And a selector unit that switches to a predetermined bit stream that has been compression-encoded and that does not reduce the compression efficiency of the bit stream stored in the bit stream memory. Processing equipment. 4. The bit stream memory has a configuration in which a predetermined bit stream corresponding to a bit stream obtained by compressing and encoding a still image with high prediction efficiency or a moving image with slow motion is stored. Moving image processing apparatus.

Images